linexactitude

Linexactitude is a neologism used in discussions about modeling and measurement to denote the degree to which a representation adheres to linear precision across its domain. As a concept, it sits between strict linearity and practical approximation, focusing on how faithfully a function or dataset behaves like a straight line over the relevant range.

One operational definition is a linexactitude score L in [0,1] for a domain D, computed as the proportion of the domain for which the function can be approximated by a linear model within a specified tolerance epsilon. If a best-fit line y = ax + b yields residuals within epsilon across most of D, L is high. In data analysis practice, R^2 or RMSE can serve as proxies; in more theory-oriented contexts, one may measure the angular deviation of the tangent so that the local linearization remains valid.

Linexactitude is particularly relevant when deciding whether a linear model is sufficient for prediction, control, or interpretation, and when comparing competing representations, such as piecewise-linear versus nonlinear models.

Origin and current use: the term is a 21st-century coinage and derives from the combination of linear and exactitude. It is not standardized or widely adopted, but appears in scholarly discussions, blogs, and some philosophy of science texts to discuss the limits and usefulness of linear approximations.

See also: linearity, linear regression, approximation theory, Taylor expansion, model adequacy.